Evaluation of treatment modalities for thoracic empyema: a cost-effectiveness analysis1

ORIGINAL ARTICLES: GENERAL THORACIC

Evaluation of Treatment Modalities for Thoracic Empyema: A Cost-Effectiveness Analysis Vinod H. Thourani, MD, Kevin M. Brady, MD, Kamal A. Mansour, MD, Joseph I. Miller, Jr, MD, and Robert B. Lee, MD Section of General Thoracic Surgery, Division of Cardiothoracic Surgery, Emory University School of Medicine, Atlanta, Georgia

Background. Empyema thoracis is treated with a multitude of therapeutic options. Optimal therapy and costcontainment requires selection of the most appropriate initial intervention. Methods. A retrospective review of treatment modalities was performed on 77 patients diagnosed with empyema thoracis from 1990 to 1997 at one institution. Mean age was 59 years (range, 21 to 90 years); 52 were men and 25 were women. Results. Sixty-five percent (50/77) were parapneumonic and 68% (52/77) were multiloculated. Treatment modalities were as follows: group 1, antibiotics only (n 5 4); group 2, primary intervention: image-directed catheter (n 5 20) or tube thoracostomy (n 5 24); and group 3, secondary intervention: decortication (n 5 17), rib resection or muscle interposition (n 5 12). Thirty-four percent (9/20 image-directed catheter and 8/24 tube thoracostomy) had failure of initial intervention. Patients undergoing decortication more often had multiloculated empyema thoracis (16 of 17) compared with those undergoing

image-directed catheters (8 of 20) or tube thoracotomy (16 of 24). Length of stay was reduced for decortication patients (17 days) compared with those having imagedirected catheters (21.8 days), failed image-directed catheters (29.7 days), or tube thoracostomies (19.6 days). Hospital charges per patient between decortication and image-directed catheter ($34,770.79 versus $37,869.41) were comparable, but charges were significantly decreased in decortication patients as compared with failed image-directed catheters ($55,609.32; p < 0.05). Conclusions. Our series revealed that early decortication has charges similar to those of primary intervention (image-directed catheter or tube thoracostomy) but is more cost-effective than failed image-directed catheter. We advocate the use of early surgical intervention as the most optimal and cost-effective initial modality for the treatment of empyema thoracis.

S

expense to society. Of the approximately 1.2 million persons affected by pneumonia in the United States, nearly 5% or 60,000 persons will develop a parapneumonic empyema [6]. Furthermore, with an aging population and an increase of intravenous drug abusers and acquired immunodeficiency syndrome-related pulmonary complications, the incidence in pneumonia is not likely to diminish. In this study, we retrospectively reviewed treatment modalities and hospital charges in patients diagnosed with ET at one institution.

ince the era of Hippocrates [1], empyema thoracis (ET) has been recognized as a debilitating disease process. Empyema thoracis has a higher frequency and is more devastating in elderly and immunocompromised populations. The most common cause of ET is parapneumonic, whereas thoracic surgical procedure, trauma, esophageal perforation, cystic fibrosis, foreign body, chest wall infections, and subdiaphragmatic abscesses are less common [2, 3]. Therapy for ET requires appropriate antibiotics, drainage of the infected space, and reexpansion of the lung [4]. However, the most appropriate and cost-effective management of ET remains controversial. Initially, most cases are treated by repeat aspiration (thoracentesis), image-directed catheters (IDC), or tube thoracostomy (CT). Traditional surgical approaches include decortication, video-assisted thoracoscopic surgery, rib resection, and open drainage with or without a musculoskeletal flap reconstruction. Despite various treatment options, ET remains associated with mortality ranging from 1% to 19% [5]. The treatment of ET poses a considerable medical

Presented at the Thirty-fourth Annual Meeting of The Society of Thoracic Surgeons, New Orleans, LA, Jan 26 –28, 1998. Address reprint requests to Dr Miller, Section of General Thoracic Surgery, 25 Prescott St, Suite 3420, Atlanta, GA 30345.

© 1998 by The Society of Thoracic Surgeons Published by Elsevier Science Inc

(Ann Thorac Surg 1998;66:1121–7) © 1998 by The Society of Thoracic Surgeons

Patients and Methods Patients A retrospective review was performed on 77 consecutive patients with a discharge diagnosis of ET at Crawford Long Hospital of Emory University between April 1990 and July 1997. Mean age was 59 6 1.9 years (range, 21 to 90 years); 52 were men and 25 were women. Empyema thoracis was confirmed in all patients by one of the This article has been selected for the open discussion forum on the STS Web site: http://www.sts.org/annals

0003-4975/98/$19.00 PII S0003-4975(98)00767-X

1122

THOURANI ET AL EMPYEMA TREATMENT: COST ANALYSIS

Ann Thorac Surg 1998;66:1121–7

Table 1. Presenting Symptoms of Patients Diagnosed With Empyema Thoracis Symptom Shortness of breath Chronic fever Chest pain Cough Acute fever Tachypnea Lethargy Anorexia Hemoptysis Asymptomatic

No. of Patients (%) 46 (60%) 41 (53%) 30 (39%) 28 (36%) 18 (23%) 16 (21%) 12 (16%) 11 (14%) 1 (1%) 1 (1%)

following three criteria: (1) aspiration of grossly purulent pleural fluid during thoracentesis, IDC, CT, or thoracotomy; (2) biochemical evidence of pleural fluid defined as pH less than 7.10, lactate dehydrogenase level greater than 1,000 IU/L, glucose level less than 40 mg/dL, protein level greater than 2.5 g, and white blood cell count (WBC) greater than 500/mL; or (3) positive pleural fluid microbiology culture or Gram stain revealing organisms. Patient charts were retrospectively reviewed for patient age, sex, chief complaint, symptoms, past medical history, first and last WBC count, length of hospital and intensive care unit stay, and mortality. Empyemas were classified by origin, microbiology culture results, and type of loculation (simple or complex). Simple or uniloculated empyemas were defined as a single fluid collection, whereas complex or multiloculated empyemas were defined as more than two fluid collections. The number of chest roentgenograms (standard and lateral decubitus) and chest computed tomography were documented. Treatment modalities were divided into three groups: group 1, antibiotics only (n 5 4); group 2, primary intervention: IDC (12F to 14F, n 5 20) or CT (n 5 24); and group 3, secondary intervention: decortication (n 5 17), Eloesser flap (n 5 8), and muscle interposition (n 5 4). Hospital charges for each patient were accrued from time of admission to Crawford Long Hospital and extended to the time of discharge, but did not incorporate outpatient follow-up visits. Charges were obtained from the Crawford Long Hospital financial department. Follow-up was obtained by chart review or telephone interviews.

Statistics Statistical analysis involved one-way analysis of variance to determine whether group-related differences occurred [7]. If significant interactions were found, Tukey’s post hoc multiple comparisons tests were applied to locate the sources of differences [7]. A p value less than 0.05 was considered significant, and mean 6 standard error of the mean are reported [7].

Results Seventy-seven patients with a discharge diagnosis of ET were treated at Crawford Long Hospital of Emory Uni-

Table 2. Causes of Empyema Thoracis Cause Parapneumonic Postthoracotomy Posttraumatic Tuberculosis-related Malignant pleural effusions Intraabdominal sepsis Esophageal-related

No. of Patients (%) 50 (65%) 12 (16%) 4 (5%) 3 (4%) 3 (4%) 3 (4%) 2 (3%)

versity from April 1990 to July 1997. The three most common presenting symptoms (Table 1) for all patients included shortness of breath (46 of 77, 60%), chronic fever (41 of 77, 53%), and chest pain (30 of 77, 39%). Sixty-five percent (50 of 77) of TE were directly related to parapneumonic effusions, and another 16% (12 of 77) occurred after thoracotomy (Table 2). In the postthoracotomy group of 12 patients, 9 (75%) were transferred to Emory University Medical Center after undergoing surgical procedures elsewhere. The majority of these patients had complications from their operation from 2 to 48 months postoperatively. The remaining 3 patients (25%) had a primary surgical procedure at Emory University Medical Center. Pneumonia developed in the residual lung of these patients from 11 to 48 months postoperatively. Ten patients (13%) had thoracotomy for lung cancer. Tuberculosis- and esophageal-related ET were uncommon. Pleural fluid cultures were obtained in all patients either during thoracentesis, primary intervention, or intraoperatively. Gram-positive cocci organisms were cultured from 42 (55%) patients diagnosed with ET (Table 3), whereas gram-negative rods were cultured from 30 (39%) patients. Acid-fast bacilli were cultured from 3 patients. Treatment modalities were divided into three groups: group 1, antibiotics only (n 5 4); group 2, primary intervention (n 5 44): IDC (n 5 20) or CT (n 5 24); and group 3, secondary intervention (29): decortication (n 5 17) or definitive drainage or reconstructive flaps (n 5 12). All four patients treated in group 1 were elderly, endstage patients who did not receive any invasive procedures. Twenty patients were initially treated with radiographic-guided IDC. Eleven patients (55%) were treated successfully with IDC, and the remaining 9 patients (45%) failed initial IDC intervention. Of these 9 patients, 3 were treated with CT (1 of the 3 patients was discharged Table 3. Microorganisms Isolated From Patients With Empyema Thoracis Type of Microorganism Gram-positive cocci Gram-negative rods Gram-positive rods Acid-fast bacilli Molds Yeast

No. of Patients (%) 42 (55%) 30 (39%) 10 (13%) 3 (4%) 1 (1%) 1 (1%)

Ann Thorac Surg 1998;66:1121–7

THOURANI ET AL EMPYEMA TREATMENT: COST ANALYSIS

1123

Table 4. Patient Group Characteristicsa

Patient Group Group 1 Antibiotics only Group 2—Initial intervention Image-directed catheters Successful Failed Tube thoracostomy Successful Failed Group 3—Secondary intervention Decortication Flap reconstruction a

aWBC (31,000/mL)

dWBC (31,000/mL)

No. of CXRs

No. of Computed Tomograms

24.1 6 2.3

23.0 6 0.6

3.3 6 1.2

1.0 6 0.4

3/4 (75%)

1/4 (25%)

14.1 6 1.0 11.6 6 0.4 17.2 6 2.4 15.1 6 1.6 16.0 6 1.9 13.3 6 1.1

8.5 6 0.9 8.0 6 1.2 9.1 6 2.0 10.7 6 1.7 11.5 6 0.4 9.7 6 4.3

16.5 6 1.7 13.1 6 2.1 20.7 6 1.0 13.6 6 1.0 12.9 6 1.3 15.0 6 0.4

3.0 6 0.2 2.8 6 0.5 3.2 6 0.6 1.7 6 0.4 1.9 6 0.2 1.3 6 0.8

8/20 (40%) 1/11 (9%) 7/9 (78%) 16/24 (67%) 10/16 (63%) 6/8 (75%)

12/20 (60%) 10/11 (91%) 2/9 (22%) 8/24 (33%) 6/16 (37%) 2/8 (25%)

12.2 6 1.9 11.3 6 2.7

10.5 6 0.8 9.6 6 1.3

14.1 6 1.0 11.8 6 2.5

1.3 6 0.4 0.6 6 0.3

16/17 (94%) 9/12 (75%)

1/17 (6%) 3/12 (25%)

Complex Effusions

Simple Effusions

Results are mean 6 standard error of the mean.

aWBC 5 admission white blood cell count;

CXRs 5 chest roentgenograms;

to home with an empyema tube), 4 underwent decortication, and 2 underwent surgical open drainage with or without muscle interposition. Twenty-four patients in group 2 with ET were treated with CT. Sixteen patients (67%) were discharged from the hospital with an empyema tube and required no further treatment. The remaining 8 patients underwent decortication (6 of 24, 25%) or other surgical procedures (2 of 24, 8%). Thirty-nine percent (9 of 20 IDC and 8 of 24 CT) of patients in group 2 failed intervention and required surgical procedures. The admission and discharge WBC counts were recorded for all patients (Table 4). Patients who were treated with antibiotics alone had a significantly higher admission WBC count than patients treated with successful IDC, decortication, and surgical flap reconstruction. Similarly, discharge WBC count for patients in group 1 was significantly increased as compared with all other treatment groups (Table 4). The number of chest roentgenograms were significantly reduced in patients in group 1 (Table 4). Although there was a significant increase in chest roentgenograms performed for failed IDC (20.7 6 1.0) as compared with patients discharged with an empyema tube (12.9 6 1.3) and patients undergoing flap reconstruction (11.8 6 2.5), there was no significant difference from patients undergoing decortication (14.1 6 1.0). The number of computed tomograms (Table 4) performed were significantly increased in patients with successful IDC (2.8 6 0.5) and failed IDC (3.2 6 0.6) as compared with patients undergoing decortication (1.3 6 0.4) or surgical flaps (0.6 6 0.3). Sixty-eight percent (52 of 77) of patients with ET had multiloculated effusions, and 33% (25 of 77) had uniloculated effusions. Seventy-five percent of patients treated in group 1 had multiloculated ET (Table 4). Twelve patients (12 of 20, 60%) treated with IDC had uniloculated effusions, and 8 patients (8 of 20, 40%) had multiloculated effusions (Table 4). Of the 11 patients who had successful IDC, 10 (91%) had uniloculated effusions and 1 patient (9%) had multiloculated effusions. In contrast, of the 9

dWBC 5 discharge white blood count.

patients in whom IDC treatment failed, only 2 (22%) had uniloculated effusions, and 7 (78%) had multiloculated effusions. Nearly all patients treated surgically, decortication (16 of 17, 94%) and surgical flap (9 of 12, 75%), had multiloculated effusions (Table 4). The length of stay for patients treated with ET was 17 6 2.5 days (range, 4 to 68 days) and the mean stay in the intensive care unit was 1.8 6 0.7 days (range, 0 to 40 days) (Table 5). Length of stay was reduced for decortication patients (17 6 2.1 days) compared with patients with an IDC (21.8 6 3.9 days) and failed initial IDC (29.7 6 2.6 days). As expected, patients who underwent flap reconstruction had the longest intensive care unit stay (5.6 6 1.4 days). There was no significant increase in intensive care unit stay for patients undergoing decortication (1.2 6 1.0 days) as compared with IDC (1.1 6 0.4 days). Ten patients (13%) died during the 7-year period (Table 5). All 4 patients (100%) who did not receive any invasive treatment died. Of the remaining 6 patients who died (2 patients had multiple treatment modalities), 2 patients had failed IDC, 3 had chest tubes, 1 had decortication, and 2 had definitive surgical flaps. Hospital charges per patient (Table 6) were comparable between IDC ($37,869.41 6 $4,308.99), successful IDC ($23,354.93 6 $2,430.23), and decortication ($34,770.79 6 $2,455.69). However, patient hospital charges for decortication were reduced as compared with failed CT ($43,168.63 6 $1,000.34) and significantly reduced compared with failed IDC ($55,609.32 6 $3,078.03).

Comment The optimal initial intervention for the treatment of ET remains controversial. As economic resources for patient health care continues to be restricted, we must amend our treatment strategies for ET to include the most cost-effective primary intervention. Studies evaluating the most economical treatment strategies for ET are lacking. In this retrospective study of 77 patients diag-

1124

THOURANI ET AL EMPYEMA TREATMENT: COST ANALYSIS

Ann Thorac Surg 1998;66:1121–7

Table 5. Patient Hospital Stay and Outcomea Patient Group Group 1 Antibiotics only Group 2—Initial intervention Image-directed catheters Successful Failed Tube thoracostomy Successful Failed Group 3—Secondary intervention Decortication Flap reconstruction a

LOS (days)

ICU Stay (days)

Mortality (%)

8.3 6 2.3

0.3 6 0.2

4/4 (100)

21.8 6 3.9 15.3 6 3.2 29.7 6 2.6 19.6 6 2.2 21.9 6 1.9 15.0 6 2.8

1.1 6 0.4 0.5 6 0.1 1.9 6 1.1 1.7 6 0.6 0.8 6 0.4 3.5 6 1.0

2/20 (10) 0/11 (0) 2/9 (22) 3/24 (13) 2/16 (13) 1/8 (13)

17.0 6 2.1 19.0 6 4.1

1.2 6 1.0 5.6 6 1.4

1/17 (6) 2/12 (17)

Results are mean 6 standard error of the mean.

ICU 5 intensive care unit;

LOS 5 length of stay.

nosed with ET at a single institution, we have reviewed hospital charges associated with medical and surgical treatment strategies for ET. Our series revealed a similar patient population as compared with reported series [4]. The most common cause of ET was parapneumonic (65%) and most effusions were multiloculated (68%). We stratified our patients into 3 groups: group 1, antibiotics only; group 2, primary intervention: IDC or CT; and group 3, secondary intervention: decortication or open surgical drainage or muscle flap interposition. Thirtyfour percent of patients failed primary intervention (IDC or CT) and required surgical intervention. Ninety-four percent of patients who underwent decortication had multiloculated effusions, compared with 40% with IDC and 67% with CT. Furthermore, 78% of patients who failed IDC had multiloculated effusions compared with 9% with successful IDC. Length of stay was reduced for decortication patients (17 days) compared with IDC (21.8 days) or CT (19.6 days). Although hospital charges per patient were comparable between IDC, CT, and decortication ($37,869.41 versus $31,412.69 versus $34,770.79), there was a decrease in charges between failed IDC or CT and decortication ($55,609.32 versus $43,168.63 versus Table 6. Hospital Chargesa Patient Group Group 1 Antibiotics only Group 2—Initial intervention Image-directed catheters Successful Failed Tube thoracostomy Successful Failed Group 3—Secondary intervention Decortication Flap reconstruction a

Hospital Charges $19,832.93 6 $2,072.57 $37,869.41 6 $4,308.99 $23,354.93 6 $2,430.23 $55,609.32 6 $3,078.03 $31,412.69 6 $3,358.08 $25,534.72 6 $4,285.64 $43,168.63 6 $1,000.34 $34,770.79 6 $2,455.69 $45,978.24 6 $6,449.29

Results are mean 6 standard error of the mean.

$34,770.79). Our study revealed that early decortication had charges similar to those of primary intervention (IDC or CT), but is more cost-effective than failed IDC. The use of early surgical intervention (decortication) for multiloculated effusions is the most cost-effective initial treatment modality for ET. Optimal effective treatment for ET requires control of infection, evacuation of pus, and reexpansion of lung. Nonsurgical treatment methods of empyema include antimicrobial regimen, daily thoracentesis and lavage, radiologically guided percutaneous catheter, intrapleural fibrinolytic debridement, or CT [8]. Appropriate systemic antibiotic therapy may reduce the progression of parapneumonic effusions and prevent the formation of an organized empyema by decreasing lung capillary leak and resulting in more rapid clearance of pleural space bacteria [9]. Antimicrobial therapy has been described to successfully resolve small uniloculated pleural infections in 81% of patients with pleural fluid positive for Gram stains or cultures [10]. Nevertheless, most physicians advocate concomitant evacuation of infected pleural space through noninvasive or surgical modalities. All of the patients in our series were placed on appropriate intravenous systemic antimicrobial agents. Of the 4 patients in our series who were treated with antibiotics alone, none survived. All 4 patients had “do not resuscitate” status and were admitted to the hospital for endstage hospice care. Radiographic-guided intervention has emerged as a timely, expedient, and accurate modality for the evacuation of pus in ET [11, 12]. Ultrasound is a useful adjunct to standard radiographic imaging when loculation of pleural fluid of empyema is suggested [13]. Ultrasound has the advantage of portability allowing the procedure to be done at the bedside of critically ill patients. However, disadvantages of ultrasound include technician dependence and the fact that marked location is performed by one individual whereas thoracentesis is performed by another individual at a different time [13]. Furthermore, there may be difficulty discerning empy-

Ann Thorac Surg 1998;66:1121–7

ema effusion from nearby consolidated lung, especially if the empyema contains air or extensive echogenic debris [14]. We do not routinely use ultrasound for the placement of IDC in the treatment of ET. The most common radiographic intervention for ET at Emory University Medical Center is the computed tomographic scanguided placement of drainage catheters. Image-directed catheters inserted using computed tomographic scan localization have been used successfully for the treatment of empyema [15, 16]. Placement of catheters and drainage should be performed as soon as the diagnosis of empyema is confirmed before formation of a visceral or parietal “peel” [13], as organization of the empyema or multiloculation of the effusion will decrease the complete drainage of the pleural effusion or empyema [14]. Success rates with IDC have been reported ranging from 57% to 90% [4, 17]. In our series, 20 patients underwent IDC as their initial intervention for ET. Fifty-five percent of patients required no further treatment. A majority of these patients (91%) presented with uniloculated effusions. However, of the 45% of patients who failed IDC treatment, 78% had multiloculated effusions and required further intervention. In agreement with previous studies [8, 14], we advocate the use of IDC for early uniloculated pleural effusions or ET, but not for multiloculated effusions. Tube thoracostomy has been used routinely and successfully for the drainage of pleural effusions or ET. The standard option for pleural space drainage is with a surgeon-performed closed-chest CT [18]. The success rates of CT in the treatment of ET has been reported at 35% to 71% [5, 19]. Failures most commonly occur as a result of placement of CT in one loculation of a multiloculated empyema, clogging of the CT with debris or thick pleural effusions, or kinking of the CT. Placement of large-bore CT is facilitated by radiographic placement (ultrasound or computed tomography) in the most dependent portion of the empyema. In our series, 24 patients received CT for treatment of ET. Sixty-seven percent were discharged with empyema tubes and did not require further treatment. Of the remaining 8 patients, all had multiloculated effusions that were not amenable to closed-chest CT or intrapleural fibrinolysis (urokinase or streptokinase) and required surgical debridement. Nonsurgical treatment modalities are adequate for most patients in whom ET has not organized. However, closed-chest drainage with antibiotics is ineffective once the empyema organizes into the fibrinopurulent stage. In these patients, surgical decortication effectively evacuates the contents of the empyema and facilitates restoration of normal respiratory function by reexpansion of the lung [20]. LeMense and colleagues [4] advocate a short trial (24 to 48 hours) of closed-chest drainage and antibiotics followed with early postprocedural radiologic imaging. If no resolution is appreciated, LeMense and colleagues practice early aggressive surgical approach with minimal morbidity and mortality. Similarly, Pothula and Krellenstein [21], Ashbaugh [22], and Van Way and associates [23] advo-

THOURANI ET AL EMPYEMA TREATMENT: COST ANALYSIS

1125

cate aggressive surgical decortication after a brief trial of closed-chest thoracostomy. Decortication remains the initial treatment of choice in patients with chronic, multiloculated empyema thoracis. In our study, we confirmed that surgical decortication can be effectively performed in patients with multiloculated effusions with excellent success rate and minimal mortality (6%). For patients who may be too ill to tolerate a major thoracotomy or with advanced multiloculated empyemas, we advocate open surgical drainage (modified Eloesser procedure) or muscle flap interposition. A variety of treatment modalities are available for the treatment of ET. The initial treatment not only depends on the clinical presentation and associated medical illnesses of the patient, but also on the organization or chronicity of the empyema. Moreover, in an era of cost-conscious medicine, the most effective primary intervention for the treatment of ET is still undetermined. Wait and coworkers [24] have shown that video-assisted thoracoscopic surgery as a primary treatment strategy for ET resulted in higher efficacy, shorter hospital stay, and less charges as compared with closed-chest drainage with fibrinolytic therapy. In our study, we conclude that early surgical treatment of multiloculated empyemas with decortication resulted in a reduction in hospital charges from failed CT and a significant reduction from failed IDC. The reduced charges of decortication may be attributed to the reduction in the length of stay and the decrease in the number of high-technological radiology procedures, including chest roentgenograms and computed tomographs. We have confirmed that surgical decortication is a safe, effective procedure for the treatment of multiloculated ET and have shown that surgical decortication is comparable in charges with the more common noninvasive treatment modalities, but is more cost-effective when these modalities fail.

References 1. Hippocrates. In: Major RH. Classic descriptions of disease. Springfield, IL: Charles C. Thomas, 1965. 2. Deschamps C, Allen MS, Trastek VF, Pairolero PC. Empyema following pulmonary resection. Chest Surg Clin N Am 1994;4:583–92. 3. Magovern CJ, Rusch VW. Parapneumonic and posttraumatic pleural space infections. Chest Surg Clin N Am 1994;4:561– 82. 4. LeMense GP, Strange C, Sahn SA. Empyema thoracis: therapeutic management and outcome. Chest 1995;107:1532–7. 5. Lemmer JH, Botham MJ, Orringer MB. Modern management of adult thoracic empyema. J Thorac Cardiovasc Surg 1985;90:849–55. 6. Light RW, Girard WM, Jenkinson SG, George RB. Parapneumonic effusions. Am J Med 1980;69:507–11. 7. Bailar JC, Mosteller F, eds. Medical uses of statistics, 2nd ed. Boston: New England Journal of Medicine Books, 1992. 8. Lee-Chiong TL. Treating empyema without surgery: how to choose the best candidates. Postgrad Med 1997;101:195–204. 9. Sahn SA. Management of complicated parapneumonic effusions. Am Rev Respir Dis 1993;148:813–7. 10. Beger HA, Morganroth ML. Immediate drainage is not required for all patients with complicated parapneumonic effusions. Chest 1990;97:731–5.

1126

THOURANI ET AL EMPYEMA TREATMENT: COST ANALYSIS

11. Merriam MA, Cronan JJ, Dorfman GS, Lambiase RE, Haas RA. Radiographically guided percutaneous catheter drainage of pleural fluid collections. AJR Am J Roentgenol 1988;151:1113–6. 12. Stavas J, van Sonneberg E, Casola G, Wittich GR. Percutaneous drainage of infected and noninfected thoracic fluid collections. J Thorac Imaging 1987;2:80–7. 13. Lee RB. Radiologic evaluation and intervention for empyema thoracis. Chest Surg Clin N Am 1996;6:439– 60. 14. Silverman SG, Mueller PR, Saini S, et al. Thoracic empyema: management with image-guided catheter drainage. Radiology 1988;169:5–9. 15. Block EFJ, Kirton OC, Windsor J, Kestner M. Guided percutaneous drainage for posttraumatic empyema thoracis. Surgery 1995;117:282–7. 16. Ulmer JL, Choplin RH, Reed JC. Image-guided catheter drainage of the infected pleural space. J Thorac Imaging 1991;6:65–73. 17. Lee KS, Im JG, Kim YH, Hwang SH, Bae WK, Lee BH. Treatment of thoracic multi-loculated empyemas with intra-

Ann Thorac Surg 1998;66:1121–7

18. 19.

20. 21.

22. 23.

24.

cavitary urokinase: a prospective study. Radiology 1991;179: 771–5. Miller JI. Empyema thoracis. Ann Thorac Surg 1990;50:343– 4. Boland GW, Lee MJ, Silverman S, Mueller R. Interventional radiology of the pleural space. Clin Radiol 1995;50: 205–14. Weissberg D, Refaely Y. Pleural empyema: 24-year experience. Ann Thorac Surg 1996;62:1026–9. Pothula V, Krellenstein DJ. Early aggressive surgical management of parapneumonic empyemas. Chest 1994;105: 832– 6. Ashbaugh DG. Empyema thoracis: factors influencing morbidity and mortality. Chest 1991;99:1162–5. Van Way C, Narrod J, Hopeman A. The role of early limited thoracotomy in the treatment of empyema. J Thorac Cardiovasc Surg 1988;96:436–9. Wait MA, Sharma S, Hohn J, Nogare AD. A randomized trial of empyema therapy. Chest 1997;111:1548–51.

DISCUSSION DR ROBERT J. KEENAN (Pittsburgh, PA): I congratulate Dr Thourani and his colleagues from Emory for continuing their work in tackling a very vexing problem in thoracic surgery. Certainly the Emory group is well-known for its work in advancing the care of these patients who still have the potential to suffer lethal consequences from this complication. I also thank Dr Thourani for the opportunity to review his manuscript in some detail so that I could better discuss it here today. Clearly this is a retrospective review encompassing an 8-year experience at one of the hospitals in the Emory system. That this is a retrospective analysis does limit the ability to formulate treatment decisions and treatment algorithms, which a prospective study might have allowed. In addition, it sounds as though a certain percentage of the patients who were treated at Crawford Long were referred there from other institutions, and these were sprinkled among the different groups. I wonder how many of these patients came to Crawford Long already with chest tubes or pigtail catheters in place. Does this restrict your ability to have managed those patients in a way that you might otherwise have managed them had they presented primarily? The time frame also encompasses the introduction of videoassisted thoracic surgery and the use of fibrinolytics, and clearly also includes a time of tremendous inflation in health care costs. Was the decision for surgical intervention/decortication predominantly influenced by the ability to do a video-assisted thoracic surgical procedure as opposed to an open thoracotomy? Were fibrinolytics used in any of the patients who had chest tubes, or catheters, to improve the efficacy of those methods of treatment? And were the hospital charges that you referred to in the report standardized to allow a better understanding of the relationship between charges in each of those three groups? In addition, clearly the pathophysiology of empyema dictates the therapy. A number of the patients who had initial drainage catheters, whether chest tubes or radiologically placed pigtails, had multiloculated effusions, and, of course, in those settings, you might suspect that that therapy would be doomed to failure. Why not just go straight to surgical drainage in those individuals? Or, if not, were those the individuals who might have benefited from fibrinolytics or the placement of a second or even a third pigtail and might that then have allowed them to be successfully treated by catheter drainage as opposed to having unsuccessful treatment? Along those lines, what was the definition of treatment failure in

your analysis? Was it based on the persistence of the symptoms or objective findings such as fever or white blood cell count? The other thing that I found interesting was 2 of the 16 patients who were successfully treated by chest tubes died. I am not sure why that should have happened. Finally, I would like to know why, particularly in this day of cost-containment (I know this is retrospective) the length of stay for successful treatment by chest tubes was so prolonged, as long as 3 weeks on average. Might you at least today advocate sending more of these patients out with empyema tubes to reduce both their potential for complications and their hospital charges? DR THOURANI: I thank Dr Keenan for his gracious remarks and his very insightful questions. As far as how many cases were transferred to Crawford Long Hospital with image-directed catheters and chest tubes and how that affected our treatment strategy; Crawford Long Hospital of the Emory University Medical Center does receive a large amount of patients who are transferred from a range of community hospitals; approximately 30% to 40% of the patients were transferred to our service. The majority of these patients who are transferred to us already have image-directed catheters or chest tubes. Moreover, for some patients who already have image-directed catheters, we (or other thoracic surgeons) are not their primary physicians and these patients inappropriately have an image-directed catheter placed within a multiloculated effusion. It is hard for us to understand the treatment modalities that our medical colleagues use in the community. By presenting reports dealing with the different treatment modalities, we may be able to better help our community physicians in appropriately treating multiloculated effusions. Were fibrinolytics used to improve the efficacy of the methods of treatment? Fibrinolytics were used on a selected basis. If the patient had residual collection of fluid after placement of the catheter, urokinase (125,000 units urokinase in 50 mL of normal saline solution plus 5 mL of 1% lidocaine) was administered within the catheter every other day for three sessions. Because our treatment protocol requires six hospital days, it is possible that the use of fibrinolytics prolongs hospital length of stay. In this study, we did not quantitate the use of fibrinolytics. As far as the hospital charges, they were evenly distributed and standardized over the 7 years that covered the span of our study. Furthermore, there were adequate amounts of patients